Adolescent physical activity participation has clear short- and long-term health benefits, yet nearly half of American youth are not physically active on a regular basis. To design successful interventions and policies for enhancing child, adolescent, and young adult physical activity, it is important to identify the determinants of different physical activity patterns across these developmental periods. This proposal is a continuation of a current research project that has already yielded promising results related to change in youth physical activity over time. Within a developmental social contextual model, the influences of demographic, person, family, peer, school, and neighborhood factors are assessed over time, as are relationships between physical activity patterns, substance use, diet, and depression. The design of the study is multilevel. Target children (N = 360) from three cohorts (10, 12, and 14 years of age) and a parent were recruited from 58 different neighborhoods. Peers of target children also are assessed. Multilevel longitudinal investigations of youth physical activity are virtually nonexistent. With its focus on individual family, school and neighborhood influences on activity over time, the proposed study is designed to provide a comprehensive, real-world view of physical activity patterns in the critical period from childhood to young adulthood. The continuation study proposes an additional 4 years of data collection, extending the current cohort-sequential design to ages 10-21. The proposed study will continue to use new statistical methods suited to the study of longitudinal and multilevel data, allowing us to examine research questions at the individual, family, school, and neighborhood levels. With 4 more years of data collection and the use of sophisticated analytic methods, this study can make a significant contribution in an understudied area, leading to better understanding of the development, correlates, mediators, constraints, and environmental context of physical activity from ages 10-21.